Modelling Ontology-based Multilayered Communities of Interest for Hybrid Recommendations

This paper describes a strategy that automatically identifies Communities of Interest (CoI) from the tastes and preferences expressed by users in personal ontology-based profiles, and presents early experiments that evaluate how these CoI can be applied to recommend annotated items combining several content-based collaborative recommendation techniques. Specifically, we have experimented with a set of synthetic profiles generated from data of the well-known IMDb and MovieLens repositories. The obtained results show the feasibility of our CoI identification and recommendation approaches.